Asger Risom
Artificial intelligence has effectively dropped the labor cost of cybercrime to zero, unleashing a wave of autonomous, machine-speed attacks that traditional corporate defenses are not equipped to stop.
For decades, the global cybersecurity landscape relied on a simple economic barrier: pulling off a sophisticated corporate breach took specialized human talent, time, and money. If your company wasn’t holding highly classified intellectual property or millions of dollars in liquid assets, you were generally safe. You just weren’t worth the human effort required to break in.
According to a new breakdown by Fortune, that protective barrier has officially collapsed. Artificial intelligence has effectively erased the labor costs of cybercrime, unleashing a relentless, automated threat environment that standard corporate defenses are simply not built to handle.
The rise of ‘vibe hacking’
The details emerging from Anthropic’s September 2026 Threat Intelligence report illustrate a terrifying shift in how breaches occur. We are no longer dealing with human hackers manually probing firewalls or typing out malicious scripts. Instead, attackers are deploying autonomous AI agents to operate at machine speed.
Anthropic researchers coined the term “vibe hacking” to describe this methodology. A human operator gives an AI model a broad objective—such as extracting data from a corporate network—and lets it loose. The AI explores the digital environment, writes its own custom scripts, tests vulnerabilities, and automatically pivots to new tactics if the first attempt fails. It is relentless trial and error without the need for human sleep or salaries.
The sheer scale of this automation is staggering. In one incident, an attacker stole a single developer token and used an AI agent to take full administrative control of a victim’s cloud environment in roughly three hours. In another breach, hackers affiliated with the “ShinyHunters” group used AI to extract more than 2,100 sets of Azure AD authentication tokens across 40 different corporate cloud environments in just 34 hours.
This automation extends well beyond basic network penetration. Anthropic tracked a Russian espionage campaign where AI agents actively monitored whether the operators’ malware was being flagged by security software. When it was detected, the AI simply rewrote and recompiled the malicious code on the fly to evade the tripwires.
An overwhelming volume of threats
This total collapse in labor costs is also supercharging social engineering. Criminals no longer need boiler rooms full of human operators to run massive phishing campaigns or catfishing operations. According to the report, a China-based app studio used Claude to actively manage 4,700 fabricated personas across a network of dating apps. In a single two-week period, this AI swarm held personalized conversations with at least 25,000 human victims to harvest in-app currency.
The fundamental reality of digital security has changed. When adversaries have an army of AI agents capable of relentlessly pounding on digital doors for fractions of a penny per hour, every single organization becomes a viable target. The era of relying on obscurity to stay safe is over, and the corporate defensive playbook needs an immediate rewrite.
